Field of the Invention
The present invention relates to an image processing technology for reducing influence of unnecessary light using a parallax image obtained by image capturing.
Description of the Related Art
When an image capturing apparatus, such as a digital still camera, captures an image, light incident on an image capturing optical system is partially reflected on a lens interface and/or a lens holder and reaches, as unnecessary light, an image capturing plane. The unnecessary light becomes ghost and flare, and a captured image has an area in which the unnecessary light is detected.
Japanese Patent Laid-Open No. (“JP”) 2014-164497 discloses a method for estimating a light source direction that causes the unnecessary light. More specifically, the light source direction is estimated using a plurality of images obtained by image captures at a plurality of different viewpoints and an environmental map representing an environment in the image capturing scene.
JP 2012-070443 discloses a method for estimating a light source direction, for acquiring information on a ghost component based on the estimation result and information of an image capturing optical system, for subtracting the ghost component from a captured image (original image), and for obtaining a ghost reduced image in which a ghost component is reduced.
The light source direction estimating method disclosed in JP 2014-164497 needs to previously set an image capturing scene and to generate the environmental map of the image capturing scene, such as an image of a semispherical image capturing scene. This configuration restricts a freedom degree of a position, an angle, and the like of an image capturing apparatus, and image capturing may be unavailable with a user preferred position, angle, etc.
The ghost component reducing method disclosed in JP 2012-070443 estimates the light source direction based on one captured image, and has difficulties in improving the estimation precision. Since the precision of specifying the area in which the unnecessary light is detected may be low, it may be difficult to sufficiently reduce this area, or this method may reduce an area other than the area in which the unnecessary light is detected. A user cannot previously recognize this fact.